CN110355761B - Rehabilitation robot control method based on joint stiffness and muscle fatigue - Google Patents

Rehabilitation robot control method based on joint stiffness and muscle fatigue Download PDF

Info

Publication number
CN110355761B
CN110355761B CN201910635595.0A CN201910635595A CN110355761B CN 110355761 B CN110355761 B CN 110355761B CN 201910635595 A CN201910635595 A CN 201910635595A CN 110355761 B CN110355761 B CN 110355761B
Authority
CN
China
Prior art keywords
muscle
joint
stiffness
model
information
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201910635595.0A
Other languages
Chinese (zh)
Other versions
CN110355761A (en
Inventor
艾青松
张从胜
刘泉
孟伟
朱承祥
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Wuhan University of Technology WUT
Original Assignee
Wuhan University of Technology WUT
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Wuhan University of Technology WUT filed Critical Wuhan University of Technology WUT
Priority to CN201910635595.0A priority Critical patent/CN110355761B/en
Publication of CN110355761A publication Critical patent/CN110355761A/en
Application granted granted Critical
Publication of CN110355761B publication Critical patent/CN110355761B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B25—HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
    • B25J—MANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
    • B25J11/00—Manipulators not otherwise provided for
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B25—HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
    • B25J—MANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
    • B25J9/00—Programme-controlled manipulators
    • B25J9/16—Programme controls
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B25—HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
    • B25J—MANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
    • B25J9/00—Programme-controlled manipulators
    • B25J9/16—Programme controls
    • B25J9/1679—Programme controls characterised by the tasks executed

Abstract

The invention discloses a rehabilitation robot control method based on joint stiffness and muscle fatigue. According to the invention, joint angle signals and surface electromyographic signals of a subject are collected, and personalized physiological parameters of muscles are identified through a genetic algorithm by combining positive and inverse dynamics principles, so that a personalized joint skeleton muscle model is established, and joint stiffness information in the motion process is calculated; calculating the median frequency of the surface myoelectric signals in the exercise process, and acquiring fatigue information of the testee by using the relative change value of the median frequency; the method adopts joint stiffness information and motion fatigue information to carry out self-adaptive adjustment on parameters of an impedance model, and simultaneously restrains stiffness and damping parameters through a saturation function, thereby realizing the self-adaptive impedance control method of the rehabilitation robot. The invention simultaneously considers the rigidity information related to the joint and the fatigue information related to the muscle, introduces the rigidity information and the fatigue information into the rehabilitation robot for control, can give consideration to the safety of the joint and the muscle in the rehabilitation training process, and realizes safe and effective rehabilitation training.

Description

Rehabilitation robot control method based on joint stiffness and muscle fatigue
Technical Field
The invention belongs to the field of rehabilitation robot control based on surface electromyographic signals, and particularly relates to a rehabilitation robot control method based on joint stiffness and muscle fatigue.
Background
The implementation of active rehabilitation control of a rehabilitation robot requires consideration of the human-machine interaction between the robot and the patient. By analyzing the biosensing information generated by the patient in the rehabilitation training process, the movement intention of the patient can be acquired, the corresponding running state of the robot is controlled accordingly, and active and safe rehabilitation training can be provided for the patient. The interactive force signal and the physiological signal are typical human-computer interaction information. The acquisition of the interaction force signal usually needs to rely on the mechanical structure of the rehabilitation robot, is not convenient and flexible enough, and cannot feed back the information of the patient in time, possibly resulting in the time delay of closed-loop control. The physiological signals can not only judge the movement intention of the patient, but also be used for evaluating the health state of limbs in the rehabilitation training process, so that the active rehabilitation control of the rehabilitation robot based on the physiological signals is widely researched. At present, active control of a rehabilitation robot based on physiological signals mostly focuses on recognition of movement intentions of human bodies, and further research on extraction of rehabilitation information such as joint stiffness, muscle fatigue states and the like is lacked.
Human joint stiffness is an important physical characteristic of a joint, which varies greatly during movement, and a rehabilitation robot needs to react to the variation during rehabilitation training to provide a safer and more comfortable control mode; the rehabilitation robot is used for carrying out rehabilitation training on a patient to recover the limb movement function of the patient to a certain extent, but the patient is easy to have muscle fatigue or fatigue deepening along with the movement, the muscle fatigue caused by over-training can cause secondary damage to the patient, so that the muscle fatigue information of the patient needs to be acquired in real time in the rehabilitation process, and the running state of the rehabilitation robot is adjusted according to the fatigue information. Based on the above, after obtaining the joint stiffness information and the exercise fatigue information of the patient, a reliable rehabilitation robot control method needs to be designed, so that the rehabilitation robot control method can safely help the patient to complete rehabilitation training.
Chinese patent CN109645962.A discloses a robot-assisted rehabilitation man-machine cooperation training method based on fatigue perception, which extracts fatigue characteristics through surface electromyographic signals and detects the fatigue state of a patient: and (4) normal/mild fatigue/severe fatigue, and switching the rehabilitation training mode according to the fatigue state and the decision controller, so as to avoid injury to the patient caused by over-training. The control method only takes the fatigue state of the patient as a control switch to switch the state of the rehabilitation training, lacks continuous feedback of fatigue information in the human-computer interaction process, cannot fully utilize the fatigue information to carry out human-computer interaction control between the two states, ignores the change of joint information in the rehabilitation training, and cannot ensure the joint safety in the rehabilitation training. Chinese patent CN106109174.A discloses an electromyographic feedback type impedance self-adaptive rehabilitation robot control method, which uses a surface electromyographic signal to identify the flexion and extension state of a joint, adaptively adjusts impedance parameters according to the joint angle and the muscle activity level, and adjusts the initial expected static balance force according to the fatigue degree to realize the self-adaptive control of the rehabilitation robot. The control method uses the healthy side of the affected side mirror image to calculate the muscle activity level to adjust the impedance parameter, is only suitable for the patient with unilateral injury, and the grading of the fatigue degree is difficult to quantitatively analyze, so that the method lacks universality. Chinese patent CN109718059.A discloses a self-adaptive control method and device for a hand rehabilitation robot, wherein the control method obtains a hand rehabilitation track through motor imagery electroencephalogram signals, and simultaneously obtains the activity of upper limbs by utilizing surface electromyogram signals so as to establish a variable impedance control model. The parameter adjustment in the control method only considers the self-adaption problem of the rehabilitation robot, and does not consider the influence of safety, so that the method has potential safety hazards; moreover, the electroencephalogram signal and the surface electromyogram signal need to be acquired simultaneously for real-time processing, and the process is difficult to implement.
Through the analysis, the design of the existing active rehabilitation controller mostly ignores the influence of the joint in the rehabilitation process according to the information of the muscle, and does not consider the safety of the joint and the muscle. In order to overcome the limitation of active control of the traditional rehabilitation robot and improve the effectiveness and safety of rehabilitation training, the invention utilizes surface electromyographic signals to acquire joint rigidity information and motion fatigue information of a human body, designs an impedance control model based on position on the basis of the joint rigidity information and the motion fatigue information, realizes self-adaptive impedance control of the rehabilitation robot based on rigidity estimation and fatigue feedback, and provides a natural, effective and safe control means for the rehabilitation training of patients.
Disclosure of Invention
In order to solve the technical problems, the invention provides a rehabilitation robot control method based on joint stiffness and muscle fatigue.
The technical scheme of the invention is a rehabilitation robot control method based on joint stiffness and muscle fatigue, which specifically comprises the following steps:
step 1: synchronously acquiring surface electromyographic signals and joint angle signals by adopting a motion acquisition and analysis system, and respectively preprocessing the surface electromyographic signals and the joint angle signals to obtain preprocessed surface electromyographic signals and actual joint angle signals;
step 2: binding muscleEstablishing a joint stiffness estimation model by using a meat activation model, a Hill muscle tendon model and a skeletal muscle geometric model, and performing EMG (electromagnetic EMG) on the preprocessed surface electromyogram signalsapRespectively inputting the actual joint angle signal theta into the joint stiffness estimation model and the joint inverse dynamics model, and identifying physiological parameters matched with the subject by using a genetic algorithm to obtain an individualized joint stiffness estimation model;
and step 3: collecting surface electromyographic signals when the joint acts in the maximum autonomous contraction, analyzing the average contribution degree of muscles of a subject in the movement process through a muscle cooperation theory, and obtaining movement fatigue information by combining the calculated median frequency of each muscle in the movement process;
and 4, step 4: collecting human-computer interaction torque in the rehabilitation training process, establishing an impedance model, setting a basic value of a stiffness parameter and a damping parameter of the rehabilitation robot, and realizing an impedance control model based on a position;
and 5: setting an expected running track of the rehabilitation robot, introducing joint stiffness information and motion fatigue information into an impedance control model based on positions to adjust stiffness parameters and damping parameters, and obtaining an actual running track.
Preferably, the synchronous collection of the surface myoelectric signals by the motion collection and analysis system in the step 1 is EMG;
in the step 1, the joint angle signals are synchronously acquired by adopting a motion acquisition and analysis system and calculated as follows:
respectively sticking mark points on the tail end of the fixed limb when the joint rotates, the joint rotation center and the tail end of the rotating limb when the joint rotates, and correspondingly recording the mark points as a point A (x)A,yA,zA) Point B (x)B,yB,zB) And point C (x)C,yC,zC) Memory for recording
Figure GDA0003622951640000031
Is a vector formed by the point B and the point A,
Figure GDA0003622951640000032
a vector formed by the point B and the point C;
acquiring joint angle signals through mark points in a motion acquisition and analysis system, and calculating joint angle signals theta through vectors formed by the mark pointsa：
Figure GDA0003622951640000033
Figure GDA0003622951640000034
Figure GDA0003622951640000035
The pretreatment of the surface myoelectric signal in the step 1 comprises the following steps:
preprocessing the surface electromyogram signal by adopting a Butterworth high-pass filter with the order of 5 and the cut-off frequency of 20Hz to reduce the influence of baseline drift and artifact noise and obtain a preprocessed surface electromyogram signal EMGap；
The joint angle signal preprocessing in the step 1 comprises the following steps:
θ＝θa-θ0
where θ is the actual joint angle signal, θ0Is the initial angular error of the marked point.
Preferably, the joint stiffness estimation model is established in combination with the muscle activation model, the Hill muscle tendon model and the skeletal muscle geometric model in step 2:
the preprocessed surface electromyogram signals EMG in the step 1 are introducedapTaking an absolute value and carrying out normalization processing to obtain a normalized surface electromyogram signal e (t) as follows:
Figure GDA0003622951640000041
wherein the EMGrea＝|EMGapI is the surface electromyography after taking the absolute valueSignals, EMGmvcProcessed surface electromyographic signal peaks, EMG, recorded during maximal voluntary contractionresSurface electromyographic signals in a resting state;
the degree of neural activation u (t) is then obtained using a second order discrete linear model as follows:
u(t)＝0.9486·e(t-de)+0.052·u(t-1)-0.000627·u(t-2)
wherein de is 80 ms. Finally, the muscle activation degree a (t) is obtained as follows:
Figure GDA0003622951640000042
introducing the actual joint angle signal theta in the step 1, and calculating the muscle length l through a muscle path equation and a personalized skeletal muscle geometric modelmtAnd the muscle force arm length d is as follows:
β＝σ+θ
Figure GDA0003622951640000043
d＝lPRlQR sinβ/lPQ
wherein P is a muscle starting point, Q is a muscle stopping point, R is a joint rotation center, and sigma is an angle between the muscle starting point, the stopping point and the rotation center in an initial state;
the muscle force F is calculated from the Hill muscle tendon model in combination with the given physiological parameters as follows:
Figure GDA0003622951640000044
wherein the content of the first and second substances,
Figure GDA0003622951640000045
is the maximum force of contraction at the optimum muscle fiber length, a is the degree of muscle activation, l is the normalized muscle fiber length, v is the normalized muscle fiber contraction velocity, FA(l) For normalized active force-muscle fiber length function，FV(v) For the normalized function of the speed of contraction of the muscle fibers of the active force, Fp(l) In order to normalize the function of passive force-muscle fiber length, alpha is called pinnate angle, and the change of output force of muscle caused by the change of pinnate angle is small in the process of movement of human muscle, so that the change of muscle force caused by the change of pinnate angle can be ignored when calculating the muscle force.
Calculating joint moment tau by using muscle force and arm length of muscle forceaThe following were used:
Figure GDA0003622951640000046
introducing the actual joint angle signal theta in the step 1, and calculating the rigidity information K of the joint rotation by using a joint rigidity estimation modelaThe following were used:
Figure GDA0003622951640000051
wherein i is the ith muscle participating in the calculation, tauaIs the joint moment.
Introducing the actual joint angle signal theta in the step 1, and calculating the reference moment tau by using a reverse dynamic modelrThe following:
Figure GDA0003622951640000052
wherein g is the gravity acceleration, and I, M, l is the moment of inertia, mass and the length of the gravity force arm of the rotating limb when the joint rotates.
Introducing the actual joint angle signal theta in the step 1, and calculating the reference rigidity K through the reference momentrThe following were used:
Figure GDA0003622951640000053
with a rigidity KaAnd a reference stiffness KrThe error between the muscle fibers is minimized to be an optimization target, and the maximum contraction force under the corresponding optimal muscle fiber length is set for each muscle
Figure GDA0003622951640000054
Optimal muscle fiber length lmoptAnd tendon length regulatory factor stThe initial values and the optimization ranges are optimized for the maximum contractility under the optimal muscle fiber length, the optimal muscle fiber length and the tendon length regulating factor through selection, crossing and variation operations of a genetic algorithm, so that optimized parameters, namely personalized physiological parameters of the testee are obtained, and a personalized joint stiffness estimation model is constructed as follows:
Figure GDA0003622951640000055
wherein the content of the first and second substances,
Figure GDA0003622951640000056
lmoptand stFor each muscle, the corresponding optimizing range is provided, H is a data point, and H is the total data point number.
Preferably, the surface electromyography signals collected during the maximum autonomous contraction of joint movements in step 3 are EMGm：
The analysis of the average contribution degree of the muscle of the subject during the exercise by the muscle synergy theory in the step 3 is as follows:
assuming that there are Q different joint motions, Root Mean Square (RMS) values are extracted from the surface electromyographic signals collected for the Q motions at maximum spontaneous contraction as muscle activity levels as follows:
Figure GDA0003622951640000061
wherein x isi,qA sequence of surface electromyographic signals representing the q-th movement, nqIndicates the length of the q-th sequence.
Calculating the ith muscle pair by adopting a muscle cooperation theoryContribution WD of the q-th actioniqThe following were used:
Figure GDA0003622951640000062
wherein N is the number of muscles, WiA muscle cooperation matrix of the ith muscle in the qth action is obtained, and the average value of the contribution degrees of all joint actions is taken as the average contribution degree c of the ith muscleiThe following were used:
Figure GDA0003622951640000063
and 3, combining the median frequency of each muscle calculated in the exercise process to obtain exercise fatigue information, wherein the exercise fatigue information is as follows:
calculating the fatigue information of the ith muscle reflected by the median frequency of the ith muscle, and acquiring the exercise fatigue information p by combining the relative change value with the average contribution degree as follows:
Figure GDA0003622951640000064
wherein p isiFatigue information for the ith muscle, p0iIs the initial fatigue information of the ith muscle.
Preferably, the establishing of the impedance model in step 4 is:
collecting the man-machine interaction torque tau, and setting the basic value of the rigidity parameter through experience, namely K0And damping parameter to obtain base value B0Because the running speed of the rehabilitation robot is generally slower than a constant speed, the acceleration term can be generally ignored, and an impedance model can be obtained as follows:
Figure GDA0003622951640000065
wherein the content of the first and second substances,
Figure GDA0003622951640000066
and q respectively represent the speed and the trajectory,
Figure GDA0003622951640000067
and q isdRespectively representing a desired speed and trajectory;
the impedance control model based on the position in the step 4 is as follows:
Figure GDA0003622951640000068
preferably, the setting of the expected track of the rehabilitation robot in step 5 is as follows:
introducing the joint stiffness estimation model in the step 1 to calculate joint stiffness information K, introducing the motion fatigue information p in the step 3, and regulating the basic values of the stiffness parameter and the damping parameter by combining the joint stiffness information K and the motion fatigue information p according to a saturation function to obtain a self-adaptively changed stiffness parameter K and a self-adaptively changed damping parameter B as follows:
Figure GDA0003622951640000071
Figure GDA0003622951640000072
wherein, B0As a base value of damping, K0As a basis value of the stiffness parameter, Bl0As a lower bound on the damping parameter, Bh0As an upper bound on the damping parameter, Kl0Is a lower bound on the stiffness parameter, Kh0The method comprises the following steps that p is an upper bound of stiffness parameters, k is stiffness information, eta is a weight factor of the fatigue information, and upsilon is a weight factor of the stiffness information;
in the step 5, the rigidity parameters and the damping parameters are adjusted in the position-based impedance control model, and the obtained actual running track is as follows:
introducing the position-based impedance control model in the step 4, and replacing the basic values of the stiffness parameter and the damping parameter with the stiffness parameter and the damping parameter which change in a self-adaptive manner to obtain the self-adaptive impedance control model as follows:
Figure GDA0003622951640000073
will be the above formula
Figure GDA0003622951640000074
Running speed of robot
Figure GDA0003622951640000075
And performing an integration operation to obtain an actual running track q.
Compared with the prior art, the invention adopts the scheme to produce the following beneficial effects:
the invention collects the human body surface electromyographic signals and the joint angle signals, establishes an individualized joint rigidity estimation model and simultaneously acquires the sports fatigue information in the rehabilitation training process. The stiffness parameter of the rehabilitation robot is adjusted by using the joint stiffness information, and the damping parameter is adjusted by using the motion fatigue information, so that the rehabilitation robot self-adaptive impedance control method based on stiffness estimation and fatigue feedback is provided. Along with the increase of the rigidity of the joint, the rigidity of the rehabilitation robot is reduced in a self-adaptive manner, so that the patient can adjust the motion posture through the effort of the patient at an angle with higher rigidity of the joint, and the condition that the overlarge rigidity of the rehabilitation robot is harmful to the joint is prevented; meanwhile, along with the deepening of fatigue, the damping parameters are also reduced in a self-adaptive manner, so that the patient can influence the motion of the rehabilitation robot more freely through own main power, and the safety of the rehabilitation training can be improved while the enthusiasm of the rehabilitation training is ensured.
Drawings
FIG. 1 is a schematic diagram of a rehabilitation robot adaptive control method based on joint stiffness and muscle fatigue according to the present invention;
FIG. 2 is a schematic diagram of the distribution of surface electromyographic signal electrodes of the ankle;
FIG. 3 illustrates the spatial location of the ankle movement gathering and analysis system markers;
fig. 4 is a schematic diagram of physiological parameter identification of the ankle flexion-extension stiffness estimation model.
2-1 is the internal calf muscle electrode position, 2-2 is the external calf muscle electrode position, 2-3 is the tibialis anterior muscle electrode position, 2-4 is the soleus muscle electrode position, 3-1 is the knee joint rotation center mark point, 3-2 is the ankle rotation center mark point, 3-3 is the toe mark point, 3-4 is the ankle flexion and extension angle, and 3-5 is the motion acquisition and analysis system mark origin.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
In the following, referring to fig. 1 to 4, the ankle flexion and extension is taken as an example to describe the specific embodiment of the present invention:
with reference to fig. 1, the ankle rehabilitation robot adaptive impedance control method based on stiffness estimation and fatigue feedback of the present invention adopts an impedance model based on position, firstly sets a reference trajectory, collects surface electromyographic signals and joint angle signals during rehabilitation training, obtains joint stiffness information and motion fatigue information, and adaptively adjusts stiffness parameters and damping parameters respectively; collecting human-computer interaction torque information by using a torque sensor, and remolding a reference track by combining an adaptive rigidity parameter and an adaptive damping parameter to generate an actual track; the actual track is input into the rehabilitation robot position controller, so that the rehabilitation robot drives the limb of the patient to move under the active movement intention of the patient, and safe and effective rehabilitation training is realized.
Step 1: synchronously acquiring surface electromyographic signals and joint angle signals by adopting a motion acquisition and analysis system, and respectively preprocessing the surface electromyographic signals and the joint angle signals to obtain preprocessed surface electromyographic signals and actual joint angle signals;
referring to fig. 2, the muscle group related to ankle flexion and extension mainly includes internal Gastrocnemius (MG), external Gastrocnemius (LG), Tibialis Anterior (TA), and Soleus (SO), and the four muscles are selected as the collection points of the surface electromyographic signals, and the specific positions are shown as 2-1, 2-2, 2-3, and 2-4 in fig. 2.
Synchronously acquiring surface myoelectric signals into EMG by adopting a motion acquisition and analysis system in the step 1;
in the step 1, the joint angle signals are synchronously acquired by adopting a motion acquisition and analysis system and calculated as follows:
referring to fig. 3, in order to calculate the ankle flexion-extension angle, a three-dimensional space is established with 3-5 (denoted as O point) as the origin of the coordinate system, so as to obtain the spatial coordinates of the three marked points. The fixed limb with the ankle rotating is the lower leg, the rotating limb is the foot, and the mark point A (x) at the tail end of the lower leg is recorded as 3-1A,yA,zA) Note 3-2 as an ankle rotation center mark point B (x)B,yB,zB) Note 3-3 as the foot end marker point C (x)C,yC,zC) Memory for recording
Figure GDA0003622951640000091
Is a vector formed by the point B and the point A,
Figure GDA0003622951640000092
the ankle flexion and extension angle 3-4 (marked as theta) is obtained through the mark points A, B and C in the motion acquisition and analysis system for the vector formed by the point B and the point Ca) Can be calculated from these two vectors:
Figure GDA0003622951640000093
Figure GDA0003622951640000094
Figure GDA0003622951640000095
the pretreatment of the surface myoelectric signal in the step 1 comprises the following steps:
preprocessing the surface electromyogram signal by adopting a Butterworth high-pass filter with the order of 5 and the cut-off frequency of 20Hz to reduce the influence of baseline drift and artifact noise and obtain a preprocessed surface electromyogram signal EMGap；
The joint angle signal preprocessing in the step 1 comprises the following steps:
θ＝θa-θ0
wherein theta is an actual joint angle signal, theta0The initial angle error of the mark point is the initial angle calculated by the mark point when the foot is vertical to the lower leg.
Step 2: establishing a joint stiffness estimation model by combining a muscle activation model, a Hill muscle tendon model and a skeletal muscle geometric model, and performing EMG (electromagnetic EMG) on the preprocessed surface electromyogram signalsapRespectively inputting the actual joint angle signal theta into the joint stiffness estimation model and the joint inverse dynamics model, and identifying physiological parameters matched with the subject by using a genetic algorithm to obtain an individualized joint stiffness estimation model;
preferably, the joint stiffness estimation model is established in combination with the muscle activation model, the Hill muscle tendon model and the skeletal muscle geometric model in step 2:
the preprocessed surface electromyogram signals EMG in the step 1 are introducedapTaking an absolute value and carrying out normalization processing to obtain a normalized surface electromyogram signal e (t) as follows:
Figure GDA0003622951640000096
wherein the EMGrea＝|EMGapI is the surface electromyographic signal after taking the absolute value, EMGmvcProcessed surface electromyographic signal peaks, EMG, recorded during maximal voluntary contractionresIs surface myoelectricity in a resting stateA signal;
the degree of neural activation u (t) is then obtained using a second order discrete linear model as follows:
u(t)＝0.9486·e(t-de)+0.052·u(t-1)-0.000627·u(t-2)
wherein de is 80 ms. Finally, the muscle activation degree a (t) is obtained as follows:
Figure GDA0003622951640000101
the muscle and the skeleton are simplified into geometric lines, a skeletal muscle geometric model is established by using a mathematical method, relevant parameters of limbs are obtained according to morphological parameters of a human body, and the length of the muscle and the length of a muscle force arm in the exercise process can be obtained by combining a muscle path evaluation equation. Introducing the actual joint angle signal theta in the step 1, and calculating the muscle length l through a muscle path equation and a personalized skeletal muscle geometric modelmtAnd the muscle force arm length d is as follows:
β＝σ+θ
Figure GDA0003622951640000102
d＝lPRlQR sinβ/lPQ
wherein P is a muscle starting point, Q is a muscle stopping point, R is a joint rotation center, and sigma is an angle between the muscle starting point, the stopping point and the rotation center in an initial state;
the muscle force F is calculated from the Hill muscle tendon model in combination with the given physiological parameters as follows:
Figure GDA0003622951640000103
wherein, the first and the second end of the pipe are connected with each other,
Figure GDA0003622951640000104
is the maximum force of contraction at the optimum muscle fiber length, a is the degree of muscle activation, l is the normalized muscle fiber length, and v isNormalized muscle fiber contraction speed, FA(l) For the normalized active force-muscle fiber length function, FV(v) For the normalized function of the speed of contraction of the muscle fibers of the active force, Fp(l) In order to normalize the function of passive force-muscle fiber length, alpha is called pinnate angle, and the change of output force of muscle caused by the change of pinnate angle is small in the process of movement of human muscle, so that the change of muscle force caused by the change of pinnate angle can be ignored when calculating the muscle force.
Calculating the muscle moment after obtaining the muscle force and the corresponding force arm, obtaining the joint moment by the sum of the moment of a plurality of muscles, taking the back extension action direction of the ankle as the positive direction, and calculating the joint moment tau of the ankle flexion and extension by using the muscle force and the length of the muscle force armaThe following were used:
Figure GDA0003622951640000111
wherein FiRepresents the muscle force produced by muscle i, diRepresenting the moment arm length of muscle i. Introducing the actual joint angle signal theta in the step 1, and calculating the rigidity information K of the joint rotation by using a joint rigidity estimation modelaThe following were used:
Figure GDA0003622951640000112
introducing the actual joint angle signal theta in the step 1, and calculating the reference moment tau by using a reverse dynamic modelrThe following:
Figure GDA0003622951640000113
wherein g is the gravity acceleration, and I, M, l is the moment of inertia, mass and the length of the gravity force arm of the rotating limb when the joint rotates.
Introducing the actual joint angle signal theta in the step 1, and calculating the reference rigidity K through the reference momentrThe following were used:
Figure GDA0003622951640000114
combined with figure 4, with stiffness KaAnd a reference stiffness KrThe minimum error between the above is the optimization target, and the maximum contraction force under the corresponding optimal muscle fiber length is respectively set for the internal calf muscle, the external calf muscle, the tibialis anterior muscle and the soleus muscle
Figure GDA0003622951640000115
Optimal muscle fiber length lmoptAnd tendon length regulatory factor stThe maximum contractility, the optimal muscle fiber length and the tendon length regulating factor of each muscle under the optimal muscle fiber length are optimized through selection, intersection and variation operations of a genetic algorithm, so that optimized parameters, namely personalized physiological parameters of the testee are obtained, and a personalized joint stiffness estimation model is constructed as follows:
Figure GDA0003622951640000116
wherein the content of the first and second substances,
Figure GDA0003622951640000117
lmoptand stFor each muscle, the corresponding optimizing range is provided, H is a data point, and H is the total data point number.
And step 3: collecting surface electromyographic signals when the joint acts in the maximum autonomous contraction, analyzing the average contribution degree of muscles of a subject in the movement process through a muscle cooperation theory, and obtaining movement fatigue information by combining the calculated median frequency of each muscle in the movement process;
step 3, collecting surface electromyographic signals of the joint during the maximum autonomous contraction of the joint action into EMGm：
The analysis of the average contribution degree of the muscle of the subject during the exercise by the muscle synergy theory in the step 3 is as follows:
the muscle cooperation theory expresses the muscle activity state as a linear combination of muscle cooperative elements and activation coefficients, and the obtained muscle cooperation matrix can express the contribution degree of the muscle to the action, as follows:
VN×T≈WN×K×HK×T
wherein, VN×TIs a surface electromyographic signal data set matrix, N is the number of selected muscles, T is the number of time samples, K is the number of synergistic elements, WN×KIs a muscle synergy matrix with K synergy elements, HK×TIs a muscle activation coefficient matrix.
The ankle flexion and extension has Q ═ 2 movements of dorsal extension and plantar flexion, and Root Mean Square (RMS) values of the surface electromyographic signals at maximum spontaneous contraction collected for these 2 movements were extracted as muscle activity levels, respectively, as follows:
Figure GDA0003622951640000121
wherein x isi,qA sequence of surface electromyographic signals representing the q-th movement, nqIndicates the length of the q-th sequence.
Taking 1 as the number of the synergistic elements of the extension of the ankle and the dorsiflexion or the plantarflexion of the foot, changing the muscle synergistic matrix into a column vector, and calculating the contribution WD of the ith muscle to the qth action by adopting a muscle synergistic theoryiqThe following were used:
Figure GDA0003622951640000122
wherein N-4 is the number of muscles, WiA muscle cooperation matrix of the ith muscle in the qth action is obtained, and the average value of the contribution degrees of all joint actions is taken as the average contribution degree c of the ith muscleiThe following:
Figure GDA0003622951640000123
and 3, combining the median frequency of each muscle calculated in the exercise process to obtain exercise fatigue information, wherein the exercise fatigue information is as follows:
the computational choice of sports fatigue is characterized by a Median Frequency (MF), calculated as follows:
Figure GDA0003622951640000124
wherein P (f) represents the power spectral density of the surface electromyogram signal, f1And f2Representing the lowest and highest frequencies of the signal bandwidth, respectively. Further available sports fatigue information is as follows:
Figure GDA0003622951640000131
wherein, MFiFatigue information for the ith muscle, MFoiAs initial fatigue information of the ith muscle, ciIs the average contribution degree of the ith muscle in the flexion and extension process.
And 4, step 4: collecting human-computer interaction torque in the rehabilitation training process, establishing an impedance model, setting the basic values of the stiffness parameter and the damping parameter of the rehabilitation robot, and realizing an impedance control model based on the position;
preferably, the establishing of the impedance model in step 4 is:
Figure GDA0003622951640000132
in the formula (15)
Figure GDA0003622951640000137
And q represent actual acceleration, velocity and trajectory respectively,
Figure GDA0003622951640000134
and q isdRespectively representing expected acceleration, speed and track, representing man-machine interaction moment by tau, acquired by a moment sensor, M0、B0And K0Representing the base values of the inertia, damping and stiffness parameters, respectively.
Collecting the man-machine interaction torque tau, and setting the basic value of the rigidity parameter through experience, namely K0And damping parameter to obtain basic value B0Because the running speed of the rehabilitation robot is generally slower than a constant speed, the acceleration term can be generally ignored, and an impedance model can be obtained as follows:
Figure GDA0003622951640000135
further, the position-based impedance control model in step 4 is:
Figure GDA0003622951640000136
and 5: setting an expected running track of the rehabilitation robot, introducing joint stiffness information and motion fatigue information into the impedance control model based on the position to adjust stiffness parameters and damping parameters, and obtaining an actual running track.
Preferably, the setting of the expected track of the rehabilitation robot in step 5 is as follows:
with reference to fig. 1, a saturation function is adopted to constrain the stiffness and damping parameters, and the stiffness of the rehabilitation robot is reduced in a self-adaptive manner along with the increase of the joint stiffness; meanwhile, as fatigue deepens, damping parameters are also reduced in a self-adaptive manner, the joint stiffness estimation model in the step 1 is introduced to calculate joint stiffness information K, the motion fatigue information p in the step 3 is introduced, and the joint stiffness information K and the motion fatigue information p are combined with a saturation function to adjust the basic values of the stiffness parameters and the damping parameters, so that the stiffness parameters K and the damping parameters B which change in a self-adaptive manner are obtained as follows:
Figure GDA0003622951640000141
Figure GDA0003622951640000142
wherein, B0As a base value of damping, K0As a basis value of the stiffness parameter, Bl0As a lower bound on the damping parameter, Bh0As an upper bound on the damping parameter, Kl0Lower bound for stiffness parameter, Kh0The method comprises the following steps that p is an upper bound of stiffness parameters, k is stiffness information, eta is a weight factor of the fatigue information, and upsilon is a weight factor of the stiffness information;
in the step 5, the rigidity parameters and the damping parameters are adjusted in the position-based impedance control model, and the obtained actual running track is as follows:
introducing the impedance control model based on the position in the step 4, and replacing the rigidity parameter and the damping parameter with the rigidity parameter and the damping parameter which change in a self-adaptive manner to obtain the self-adaptive impedance control model as follows:
Figure GDA0003622951640000143
will be the above formula
Figure GDA0003622951640000144
Running speed of robot
Figure GDA0003622951640000145
And performing an integration operation to obtain an actual running track q.
The basic values, the upper bound and the lower bound of the damping parameters and the rigidity parameters, and the weight factors of the fatigue information and the rigidity information can be set according to the experience of a specific experiment platform and experimenters.
It should be understood that parts of the specification not set forth in detail are well within the prior art.
It should be understood that the above description of the preferred embodiments is given for clarity and not for any purpose of limitation, and that various changes, substitutions and alterations can be made herein without departing from the spirit and scope of the invention as defined by the appended claims.

Claims (2)

1. A rehabilitation robot control method based on joint stiffness and muscle fatigue is characterized by comprising the following steps:
step 1: synchronously acquiring surface electromyographic signals and joint angle signals by adopting a motion acquisition and analysis system, and respectively preprocessing the surface electromyographic signals and the joint angle signals to obtain preprocessed surface electromyographic signals and actual joint angle signals;
step 2: establishing a joint stiffness estimation model by combining a muscle activation model, a Hill muscle tendon model and a skeletal muscle geometric model, and performing EMG (electromagnetic EMG) on the preprocessed surface electromyogram signalsapRespectively inputting the actual joint angle signal theta into the joint stiffness estimation model and the joint inverse dynamics model, and identifying physiological parameters matched with the subject by using a genetic algorithm to obtain an individualized joint stiffness estimation model;
and step 3: collecting surface electromyographic signals when the joint acts in the maximum autonomous contraction, analyzing the average contribution degree of muscles of a subject in the movement process through a muscle cooperation theory, and obtaining movement fatigue information by combining the calculated median frequency of each muscle in the movement process;
and 4, step 4: collecting human-computer interaction torque in the rehabilitation training process, establishing an impedance model, setting a basic value of a stiffness parameter and a damping parameter of the rehabilitation robot, and realizing an impedance control model based on a position;
and 5: setting an expected running track of the rehabilitation robot, introducing joint stiffness information and motion fatigue information into a position-based impedance control model to adjust stiffness parameters and damping parameters, and obtaining an actual running track;
and 2, establishing a joint stiffness estimation model by combining the muscle activation model, the Hill muscle tendon model and the skeletal muscle geometric model:
introducing the pretreated surface in step 1Electromyographic signals EMGapTaking an absolute value and carrying out normalization processing to obtain a normalized surface electromyogram signal e (t) as follows:
Figure FDA0003622951630000011
wherein the EMGrea＝|EMGapI is the surface electromyographic signal after taking the absolute value, EMGmvcProcessed surface electromyographic signal peaks, EMG, recorded during maximal voluntary contractionresThe surface electromyogram signal is in a resting state;
the degree of neural activation u (t) is then obtained using a second order discrete linear model as follows:
u(t)＝0.9486·e(t-de)+0.052·u(t-1)-0.000627·u(t-2)
wherein de is 80 ms; finally, the muscle activation degree a (t) is obtained as follows:
Figure FDA0003622951630000021
introducing the actual joint angle signal theta in the step 1, and calculating the muscle length l through a muscle path equation and a skeletal muscle geometric modelmtAnd the muscle force arm length d is as follows:
β＝σ+θ
Figure FDA0003622951630000022
d＝lPRlQRsinβ/lPQ
wherein P is a muscle starting point, Q is a muscle stopping point, R is a joint rotation center, and sigma is an angle between the muscle starting point, the stopping point and the rotation center in an initial state;
the muscle force F is calculated from the Hill muscle tendon model in combination with the given physiological parameters as follows:
Figure FDA0003622951630000023
wherein the content of the first and second substances,
Figure FDA0003622951630000024
is the maximum force of contraction at the optimum muscle fiber length, a is the degree of muscle activation, l is the normalized muscle fiber length, v is the normalized muscle fiber contraction velocity, FA(l) For the normalized active force-muscle fiber length function, FV(v) For the normalized function of the speed of contraction of the muscle fibers of the active force, Fp(l) As a normalized passive power-muscle fiber length function, α is called the pinnate angle;
calculating joint moment tau by using muscle force and arm length of muscle forceaThe following were used:
Figure FDA0003622951630000025
introducing the actual joint angle signal theta in the step 1, and calculating the rigidity K of the joint rotation by using the derivative of the joint torque on the actual joint angle signalaThe following were used:
Figure FDA0003622951630000026
wherein i is the ith muscle participating in the calculation, tauaIs the joint moment;
introducing the actual joint angle signal theta in the step 1, and calculating the reference moment tau by using a reverse dynamic modelrThe following were used:
Figure FDA0003622951630000027
wherein g is the gravity acceleration, and I, M, l is the moment of inertia, mass and the length of the gravity force arm of the rotating limb when the joint rotates;
introducing the actual joint angle signal theta in the step 1,calculating a reference stiffness K from a reference momentrThe following were used:
Figure FDA0003622951630000031
with a rigidity KaAnd a reference stiffness KrThe error between the muscle fibers is minimized to be an optimization target, and the maximum contraction force under the corresponding optimal muscle fiber length is set for each muscle
Figure FDA0003622951630000032
Optimum muscle fiber length lmoptAnd tendon length regulatory factor stThe initial values and the optimization ranges are optimized for the maximum contraction force, the optimal muscle fiber length and the tendon length regulating factor under the optimal muscle fiber length through selection, crossing and variation operations of a genetic algorithm, so that optimized parameters, namely personalized physiological parameters of a testee are obtained, and a personalized joint stiffness estimation model is constructed as follows:
Figure FDA0003622951630000033
wherein the content of the first and second substances,
Figure FDA0003622951630000034
lmoptand stFor each muscle, a corresponding optimizing range is provided, H is a data point, and H is the total data point number;
step 3, collecting surface electromyographic signals of the joint during the maximum autonomous contraction of the joint action into EMGm：
The analysis of the average contribution degree of the muscle of the subject during the exercise by the muscle synergy theory in the step 3 is as follows:
assuming that Q different joint motions are provided, root mean square values RMS of surface electromyographic signals collected by the Q motions and used during maximum autonomous contraction are respectively extracted as the muscle activity levels as follows:
Figure FDA0003622951630000035
wherein x isi,qA sequence of surface electromyographic signals representing the q-th movement, nqRepresents the length of the q sequence;
calculating the contribution WD of the ith muscle to the qth action by adopting a muscle cooperation theoryi，qThe following were used:
Figure FDA0003622951630000036
wherein N is the number of muscles, Wi,qA muscle cooperation matrix of the ith muscle in the qth action is obtained, and the average value of the contribution degrees of all joint actions is taken as the average contribution degree c of the ith muscleiThe following were used:
Figure FDA0003622951630000041
and 3, combining the median frequency of each muscle calculated in the exercise process to obtain exercise fatigue information, wherein the exercise fatigue information is as follows:
calculating the fatigue information of the ith muscle reflected by the median frequency of the ith muscle, and acquiring the exercise fatigue information p by combining the relative change value of the fatigue information of the ith muscle reflected by the median frequency of the ith muscle with the average contribution degree as follows:
Figure FDA0003622951630000042
wherein p isiFatigue information for the ith muscle, p0iInitial fatigue information of the ith muscle;
the impedance model establishment in the step 4 is as follows:
collecting the man-machine interaction torque tau, and setting the basic value of the rigidity parameter through experience, namely K0And the base value of the damping parameter, i.e. B0To obtain an impedanceThe model is as follows:
Figure FDA0003622951630000043
wherein the content of the first and second substances,
Figure FDA0003622951630000044
and q respectively represent the speed and the trajectory,
Figure FDA0003622951630000045
and q isdRespectively representing a desired speed and trajectory;
the impedance control model based on the position in the step 4 is as follows:
Figure FDA0003622951630000046
in the step 5, introducing the joint stiffness information and the motion fatigue information into the impedance control model based on the position to adjust the stiffness parameter and the damping parameter as follows:
introducing the individualized joint stiffness estimation model in the step 2 to calculate joint stiffness information K, introducing the motion fatigue information p in the step 3, and regulating the basic values of the stiffness parameter and the damping parameter by combining the joint stiffness information K and the motion fatigue information p according to a saturation function to obtain a self-adaptively changed stiffness parameter K and a self-adaptively changed damping parameter B as follows:
Figure FDA0003622951630000047
Figure FDA0003622951630000051
wherein, B0Is a base value of a damping parameter, K0As a basis value of the stiffness parameter, Bl0As a lower bound on the damping parameter, Bh0To hinderUpper bound of the damping parameter, Kl0Is a lower bound on the stiffness parameter, Kh0The method comprises the following steps that p is an upper bound of stiffness parameters, k is stiffness information, eta is a weight factor of the fatigue information, and upsilon is a weight factor of the stiffness information;
in step 5, the actual running track is calculated by the following specific method:
introducing the impedance control model based on the position in the step 4, and replacing the rigidity parameter and the damping parameter with the rigidity parameter and the damping parameter which change in a self-adaptive manner to obtain the self-adaptive impedance control model as follows:
Figure FDA0003622951630000052
will be the above formula
Figure FDA0003622951630000053
Speed of robot in (1)
Figure FDA0003622951630000054
And performing integral operation to obtain an actual running track q.
2. The rehabilitation robot control method based on joint stiffness and muscle fatigue according to claim 1, characterized in that: synchronously acquiring surface myoelectric signals into EMG by adopting a motion acquisition and analysis system in the step 1;
in the step 1, the joint angle signals are synchronously acquired by adopting a motion acquisition and analysis system and calculated as follows:
respectively sticking mark points on the tail end of the fixed limb when the joint rotates, the joint rotation center and the tail end of the rotating limb when the joint rotates, and correspondingly recording the mark points as a point A (x)A,yA,zA) Point B (x)B,yB,zB) And point C (x)C,yC,zC) Memory for recording
Figure FDA0003622951630000055
Is the vector formed by point B and point a,
Figure FDA0003622951630000056
a vector formed by a point B and a point C;
acquiring joint angle signals through mark points in a motion acquisition and analysis system, and calculating joint angle signals theta through vectors formed by the mark pointsa：
Figure FDA0003622951630000057
Figure FDA0003622951630000058
Figure FDA0003622951630000061
The pretreatment of the surface myoelectric signal in the step 1 comprises the following steps:
preprocessing the surface electromyogram signal by adopting a Butterworth high-pass filter with the order of 5 and the cut-off frequency of 20Hz to reduce the influence of baseline drift and artifact noise and obtain a preprocessed surface electromyogram signal EMGap；
The joint angle signal preprocessing in the step 1 comprises the following steps:
θ＝θa-θ0
where θ is the actual joint angle signal, θ0Is the initial angular error of the marked point.
CN201910635595.0A 2019-07-15 2019-07-15 Rehabilitation robot control method based on joint stiffness and muscle fatigue Active CN110355761B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910635595.0A CN110355761B (en) 2019-07-15 2019-07-15 Rehabilitation robot control method based on joint stiffness and muscle fatigue

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910635595.0A CN110355761B (en) 2019-07-15 2019-07-15 Rehabilitation robot control method based on joint stiffness and muscle fatigue

Publications (2)

Publication Number Publication Date
CN110355761A CN110355761A (en) 2019-10-22
CN110355761B true CN110355761B (en) 2022-06-14

Family

ID=68219281

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910635595.0A Active CN110355761B (en) 2019-07-15 2019-07-15 Rehabilitation robot control method based on joint stiffness and muscle fatigue

Country Status (1)

Country Link
CN (1) CN110355761B (en)

Families Citing this family (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111897415A (en) * 2020-06-22 2020-11-06 东南大学 Virtual artificial hand compliance visual control method based on electromyographic signals and variable stiffness control
CN111904795B (en) * 2020-08-28 2022-08-26 中山大学 Variable impedance control method for rehabilitation robot combined with trajectory planning
CN112364785B (en) * 2020-11-13 2023-07-25 中移雄安信息通信科技有限公司 Exercise training guiding method, device, equipment and computer storage medium
CN113084813B (en) * 2021-04-13 2022-05-03 中国科学院自动化研究所 Robot motion control method for optimally constructing constraint force field based on muscle parameters
CN113662533B (en) * 2021-07-15 2022-04-22 华中科技大学 Joint rehabilitation movement monitoring and management system and use method
CN113627002B (en) * 2021-07-30 2023-12-29 华中科技大学 Distributed force measuring point optimization method based on man-machine coupling dynamics model and application
CN114700959B (en) * 2021-12-01 2024-01-30 宁波慈溪生物医学工程研究所 Mechanical arm mirror image impedance control method and mirror image mechanical arm equipment
CN114224689A (en) * 2021-12-20 2022-03-25 广州中医药大学(广州中医药研究院) Lower limb rehabilitation exoskeleton device and control method thereof
CN114800440B (en) * 2022-03-09 2023-07-25 东南大学 Outer limb robot auxiliary supporting method based on variable rigidity
CN114700942A (en) * 2022-03-25 2022-07-05 宁波慈溪生物医学工程研究所 Upper limb robot optimization method and device and upper limb robot
CN114931390B (en) * 2022-05-06 2023-06-13 电子科技大学 Muscle strength estimation method based on fatigue analysis
CN114948591B (en) * 2022-05-12 2023-08-01 中山大学 Control method and device of lower limb rehabilitation robot and robot
CN115708758A (en) * 2022-11-19 2023-02-24 哈尔滨理工大学 Upper limb rehabilitation mode and training method based on flexible mechanical arm and human body myoelectric signal
CN117100293B (en) * 2023-10-25 2024-02-06 武汉理工大学 Muscle fatigue detection method and system based on multidimensional feature fusion network

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105288933A (en) * 2015-11-20 2016-02-03 武汉理工大学 Self-adaptation training control method of parallel lower limb rehabilitation robot and rehabilitation robot
CN107854284A (en) * 2017-12-13 2018-03-30 华中科技大学 A kind of ankle-joint ectoskeleton based on elastic element rigidity handover mechanism
CN109262618A (en) * 2018-12-12 2019-01-25 武汉理工大学 Upper limb multi-joint synchronous proportional myoelectric control method and system based on muscle collaboration
KR20210098256A (en) * 2020-01-31 2021-08-10 대전대학교 산학협력단 Neuromuscular electrical stimulator and stimulation method using emg feedback

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP2801389B1 (en) * 2013-05-08 2022-06-08 Consejo Superior De Investigaciones Científicas (CSIC) Neuroprosthetic device for monitoring and suppression of pathological tremors through neurostimulation of the afferent pathways
EP3649633A4 (en) * 2017-07-06 2021-03-10 Icuemotion LLC Systems and methods for data-driven movement skill training

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105288933A (en) * 2015-11-20 2016-02-03 武汉理工大学 Self-adaptation training control method of parallel lower limb rehabilitation robot and rehabilitation robot
CN107854284A (en) * 2017-12-13 2018-03-30 华中科技大学 A kind of ankle-joint ectoskeleton based on elastic element rigidity handover mechanism
CN109262618A (en) * 2018-12-12 2019-01-25 武汉理工大学 Upper limb multi-joint synchronous proportional myoelectric control method and system based on muscle collaboration
KR20210098256A (en) * 2020-01-31 2021-08-10 대전대학교 산학협력단 Neuromuscular electrical stimulator and stimulation method using emg feedback

Also Published As

Publication number Publication date
CN110355761A (en) 2019-10-22

Similar Documents

Publication Publication Date Title
CN110355761B (en) Rehabilitation robot control method based on joint stiffness and muscle fatigue
CN108785997B (en) Compliance control method of lower limb rehabilitation robot based on variable admittance
CN105796286B (en) Use the lower limb exoskeleton robot control method of air bag sensor
CN109718059B (en) Self-adaptive control method and device for hand rehabilitation robot
JPH01502005A (en) Apparatus and method for analyzing coordination of movements
Villa-Parra et al. Towards a robotic knee exoskeleton control based on human motion intention through EEG and sEMGsignals
WO2023206833A1 (en) Wrist rehabilitation training system based on muscle synergy and variable stiffness impedance control
CN107440887A (en) Complete bionical class brain intelligent hand electric mechanical ectoskeleton and its complex control system
CN111408042B (en) Functional electrical stimulation and lower limb exoskeleton intelligent distribution method, device, storage medium and system
US11738194B2 (en) Closed loop computer-brain interface device
Farina et al. Surface Electromyography for MAN‐Machine Interfacing in Rehabilitation Technologies
CN110339024A (en) Lower limb exoskeleton robot and its real-time gait switching method and storage device
Duvinage et al. Control of a lower limb active prosthesis with eye movement sequences
Oyama et al. Biomechanical reconstruction using the tacit learning system: intuitive control of prosthetic hand rotation
Gregory et al. Intent prediction of multi-axial ankle motion using limited EMG signals
Mu et al. Application of artificial intelligence in rehabilitation assessment
WO2012107096A1 (en) Method for determining an artificial periodic patterned signal
CN111243705A (en) Self-adaptation VR mirror image training system
CN209253488U (en) A kind of bionical class brain intelligent hand electric mechanical ectoskeleton and its control system entirely
Tong et al. BP-AR-based human joint angle estimation using multi-channel sEMG
Mohamed Modeling and simulation of transfemoral amputee gait
CN114617745A (en) Lower limb rehabilitation robot training control method and system
Popović et al. Central nervous system lesions leading to disability
Mercado et al. Hybrid BCI approach to control an artificial tibio-femoral joint
CN117398264A (en) Lower limb rehabilitation system capable of automatically switching active control modes and control method

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant